Back in February 2026, nearly 6,000 CEOs and senior executives across the US, UK, Germany, and Australia were asked a straightforward question. Has AI impacted productivity or employment at your company over the past 3 years? 90% said no. No productivity gains, no employment changes, nothing really measurable. Meanwhile, in the same 3-month window, the tech industry laid off 80,000 workers, and nearly half of those cuts were explicitly blamed on AI. So, which is it? Is AI transforming work, or is it doing absolutely nothing? The answer is messier and more revealing than either narrative admits. I'm L, I have a PhD in computer science, and I analyze AI developments to understand what's actually happening beneath all of the hype. First, let's understand what these executives are actually saying. Their survey wasn't asking about AI in general. It asked about impact at their own firms. Around 70% of companies reported actively using AI, but the CEOs themselves, they're spending an average of 1.5 hours per week with the technology. A quarter of them don't use it at all. Just think about that for a second. The people making billion-dollar AI investment decisions are barely touching the tools. They're not really seeing productivity jumps for themselves. They're not cutting head count because AI made roles obsolete, and yet companies spent over $250 billion on AI in 2024 alone, and that number keeps climbing. This disconnect has an actual name. In 1987, Nobel laureate Robert Solow said about computers, I'm quoting, "You can see the computer age everywhere but in the productivity statistics." 40 years later, economists are dusting off the same observation for AI. Apollo's chief economist wrote, I'm quoting, "AI is everywhere except in the incoming macroeconomic data." The pattern appears to be identical. Massive hype, enormous investment, very little measurable economic impact. The question isn't whether this resolves the way computers did with eventual productivity gains after a lag, or whether this is something else entirely.

Here's where it gets actually complicated. Some data does show gains. The Federal Reserve Bank of St. Louis found 1.9% excess productivity growth since ChatGPT launched in late 2022. MIT's Daron Acemoglu projected a more modest 0.5% increase over the next decade. But when executives were asked what they expect over the next 3 years, the numbers jumped. 1.4 productivity increase, 0.8% output growth. The paradox isn't that AI doesn't work, is that the people deploying it can't point to where it's working yet, but they're convinced that it will do it soon. That gap between current reality and future expectation is exactly where all of the chaos lives.

Now, let's talk about the layoffs, because this is where the story fractures. In Q1 2026 alone, 78,000 tech workers lost their jobs. Nearly 48% of those cuts were attributed to AI. Block, the company behind Cash App, cut 40% of its workforce, roughly 4,000 people, with CEO Jack Dorsey stating explicitly, and I'm quoting here, "This is not driven by financial difficulty, but by the growing capability of AI tools." Oracle reportedly laid off between 10,000 and 30,000 employees in a single day, framing it as a part of a shift toward AI infrastructure investment. These are not struggling companies at all. Many reported strong earnings right before these cuts. But here's the problem. Babak Hodjat, Cognizant's chief AI officer, told reporters, I'm quoting him here, "I don't know if they are directly related to actual productivity gains. Sometimes AI becomes the scapegoat from a financial perspective, like when a company hired too many, or they want to resize, and it gets blamed on AI." OpenAI's Sam Altman said the same thing, I'm quoting again here, "There's some AI washing where people are blaming AI for layoffs that they would otherwise do." In normal words, some of these cuts are real AI displacement, some are cost-cutting dressed up as innovation. The companies themselves are not really being clear about which is which. Maybe they don't even know. The clearest example of this mess is Klarna, actually. In 2024, the fintech company replaced 700 customer service workers with an OpenAI-powered chatbot. The CEO bragged publicly about the move. By mid-2025, customer satisfaction had cratered. The AI handled volume, yes, but couldn't manage nuance, empathy, or complex problem-solving. Klarna quietly started rehiring humans. "We went too far," they admitted. Research found that 55% of companies that executed AI-driven layoffs now appear to regret it. The cost savings evaporated when you factored in rehiring, retraining, and the reputational damage from angry customers.

So, what is actually happening? Well, multiple things at the same time. Some roles are being automated, coding assistance, content generation, basic data processing, but many companies are cutting jobs for normal traditional reasons, post-pandemic overcorrection, funding AI infrastructure, margin pressure, and using AI as a cover because it sounds forward-thinking. Amazon, Meta, Google, and Microsoft are expected to invest a combined $650 billion in AI infrastructure. That capital has to come from somewhere. Payroll is the largest controllable cost. The narrative that AI replaced these workers is just cleaner than we needed cash to build data centers. Meanwhile, the productivity data keeps contradicting the hype. A PwC survey of 4,454 CEOs found that 56% of them got nothing out of their AI investment. Only 12% reported both revenue growth and cost reduction from AI. Deloitte found that 66% report productivity gains, but only 20% see revenue growth. What that tells you is AI might be making tasks faster, but it's not translating into measurable business outcomes yet. The time saved often gets eaten by reviewing AI output, fixing errors, or managing the integration itself. Research found that 37 to 40% of time supposedly saved by AI gets consumed correcting what it produces. There's also this. Worker confidence in AI dropped 18% in 2025, even as usage increased 13%. People are using the tools more, but trusting them less. That is a psychological dynamic that doesn't show up in productivity stats, but matters enormously for adoption. If your team treats AI output as something to verify rather than trust, you haven't really automated the work, you've added a verification layer.

So, what happens next? Economists are split on this one. Some see this as a classic J-curve, short-term disruption followed by exponential gains once companies figure out how to restructure workflows around AI rather than bolting it onto existing processes. Others think the gains will be real, but modest, 0.5% to 2% productivity growth over a decade, not the 40% performance jump that early MIT studies were claiming. The gap between those two futures is enormous, and we don't really know which one we're in for another year or two. What is clear right now, however, is this. AI is being used to justify decisions that would have happened anyway, and that creates a fog where genuine displacement and opportunistic cost-cutting become impossible to separate. Some people are losing jobs because AI generally replaced their function, sure. Others are losing jobs because their company needed to fund AI infrastructure or cut costs, and blaming AI sounds better than admitting you overhired back in 2021. Both are happening at the same time. The companies doing the cutting aren't being transparent about which is which, and the workers caught in the middle are sadly left trying to figure out if retraining matters, or if the entire structure is shifting under them. The uncomfortable reality is that we're in a phase where the economic data lags the narrative by years. Companies are making bets on a future they cannot really measure yet. Some will be right, some, like Klarna, will overshoot and quietly walk it back. The ones who get it right will be the ones who treat AI as a tool for augmenting human work, not replacing it wholesale. The ones who get it wrong will discover that rehiring is very expensive, and trust is even harder to rebuild. Right now, watch the video on your screen about how France is investing in Mistral and properly challenging US AI dominance through a different paradigm and architecture altogether. Thank you all so much for watching. Subscribe, and I'll see you all on the next one.